Editorial

Placebos in practice

BMJ 2004; 329 doi: http://dx.doi.org/10.1136/bmj.329.7472.927 (Published 21 October 2004) Cite this as: BMJ 2004;329:927
  1. David Spiegel (dspiegel{at}stanford.edu), Willson professor
  1. Stanford Medical Center, Stanford, CA 94305-5718, USA

    Doctors use them, they work in some conditions, but we don't know how they work

    Placebo comes from the Latin for “I will please.” Pleasing a patient would seem to be a good thing to do. Yet considerable controversy exists about the use of a biologically inert or irrelevant substance with therapeutic intent. Nitsan and Lichtenberg show in this issue (p 944) that placebos are often used in modern medicine.1 Their survey of 89 doctors and nurses providing hospital based and ambulatory care in Israel found that 60% used placebos in their practice, most often (43%) to fend off an “unjustified” demand for medication, to calm a patient (38%), as an analgesic (38%), or, more problematically, as a diagnostic tool (28%). The paper makes it clear that the placebo pleases modern doctors. Should it? If the placebo effect is real, is it right to use it?

    Most of medicine used placebos at one time. Medicine in the 20th century was supposed to end this. We would use only scientifically proved active pharmacological and surgical interventions. Yet only about half of medical treatments are supported by evidence.2

    A conference held at the National Institutes of Health in the United States in 2002 reviewed the evidence and concluded that we needed more science, not fewer placebos.3 It called for research on brain and body pathways that mediate placebo effects, and for optimising the use of the placebo phenomenon while attending to ethical and practical concerns about it. The response to placebos varies widely across different cultures, but the myth that placebo responders are uneducated, unintelligent, free of serious medical illness, or histrionic was not supported by current research.w1 The conference also noted the importance of the nocebo effect—negative expectations can produce negative results.w2 In the 19th century, for example, tomatoes were believed to be poisonous, and many people were treated in hospitals for symptoms of tomato poisoning.w3

    Doubts have been raised about the usefulness of the placebo in conditions other than pain. A recent meta-analysis and related Cochrane report found little evidence that using a placebo improved symptoms, with the exception of pain relief.4 5 This meta-analysis had numerous problems.6 It lumped together 40 heterogeneous outcome assessments and further restricted the power to detect an effect by dichotomising them as improved or not, rather than treating these outcomes as continuous variables. The one analysis that did treat outcome continuously found a significant effect—reduction in pain. What this meta-analysis really showed is not that the placebo doesn't cure anything, but rather that it does not cure everything.

    Despite such scientific scepticism, the placebo phenomenon seems to influence the behaviour of patients and doctors. The growth of patients' interest in integrative or holistic medicine in the past decade is perhaps a reaction to the growth of evidence based medicine.7 8 Most placebos are relatively harmless. Modern medicine involves treatments, such as surgery, chemotherapy, and bone marrow transplantation, that are effective but also toxic. Many patients may choose integrative medicine as a kinder and gentler treatment that harnesses rather than eschews the placebo effect and engages them as participants in their care, especially in the treatment of chronic problems such as anxiety and pain that are often not well managed in medicine.w4

    That an idea, feeling, or relationship can have a real effect on the body is now established. Scientific domains such as psychoneuroimmunology and psychoneuroendocrinology are helping us to understand mechanisms whereby belief in benefit might affect resistance to disease, for example as a form of stress reduction with physiological consequences. Considerable evidence indicates that depression, for example, affects outcome in heart disease and cancer.9 10

    In the survey by Nitsan and Lichtenberg, few doctors (15%) used placebos without deception.1 However, deception is not a necessary component of the placebo response. Many have justifiable ethical concerns about deliberately deceiving patients regarding the nature of their treatment. But perhaps therapeutic intent and expectation are sufficient to mobilise healing in patients.w5

    The use of the placebo as a diagnostic tool by 28% of respondents is more troubling, as the authors note. That a patient gets pain relief from a placebo does not imply that the pain is not real or organic in origin. Ample evidence shows that psychological interventions such as hypnosis can alter the perception of pain dramatically. Beecher noted that grievously wounded soldiers on the Anzio beachhead required less analgesia than less seriously injured patients in Boston.11 Their intense desire to survive overrode their focus on the pain of their injuries. Psychological factors such as redirected attention, distraction, and changes in perceived meaning of real pain can alter its intensity. Thus the use of the placebo for “diagnosis” of whether or not pain is real is misguided.

    The placebo effect, thought of as the result of the inert pill, can be better understood as an effect of the relationship between doctor and patient.12 Adding the doctor's caring to medical care affects the patient's experience of treatment, reduces pain, and may affect outcome. This survey makes it clear that doctors continue to use placebos, and most think they help. We cannot afford to dispense with any treatment that works, even if we are not certain how it does.

    Footnotes

    • Papers p 944

    • Embedded ImageReferences w1 to w5 are on bmj.com

    • Competing interests None declared.

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.